Advanced EMI shielding and thermal management materials are becoming critical for today's microelectronic, photonic, and photovoltaic systems. These systems require shielding against EMI from external sources, and these systems can be sources of electromagnetic interference to other sensitive electronic devices and must be shielded. Materials for EMI shielding applications must be electrically conducting.
Further, as new and more powerful chip designs and light-emitting diode (LED) systems are introduced, they consume more power and generate more heat. This has made thermal management a crucial issue in today's high performance systems. Systems ranging from active electronically scanned radar arrays, web servers, large battery packs for personal consumer electronics, wide-screen displays, and solid-state lighting devices all require high thermal conductivity materials that can dissipate heat more efficiently. On the other hand, the devices are designed and fabricated to become increasingly smaller, thinner, lighter, and tighter. This further increases the difficulty of thermal dissipation. Actually, thermal management challenges are now widely recognized as the key barriers to industry's ability to provide continued improvements in device and system performance.
Heat sinks are components that facilitate heat dissipation from the surface of a heat source, such as a CPU or battery in a computing device, to a cooler environment, such as ambient air. Typically, heat transfer between a solid surface and the air is the least efficient within the system, and the solid-air interface thus represents the greatest barrier for heat dissipation. A heat sink is designed to enhance the heat transfer efficiency between a heat source and the air mainly through increased heat sink surface area that is in direct contact with the air. This design enables a faster heat dissipation rate and thus lowers the device operating temperature.
Materials for thermal management applications (e.g. as a heat sink) must be thermally conducting. Typically, heat sinks are made from a metal, especially copper or aluminum, due to the ability of metal to readily transfer heat across its entire structure. Cu and Al heat sinks are formed with fins or other structures to increase the surface area of the heat sink, often with air being forced across or through the fins to facilitate heat dissipation of heat to the air. However, there are several major drawbacks or limitations associated with the use of metallic heat sinks. One drawback relates to the relatively low thermal conductivity of a metal (<400 W/mK for Cu and 80-200 W/mK for Al alloy). In addition, the use of copper or aluminum heat sinks can present a problem because of the weight of the metal, particularly when the heating area is significantly smaller than that of the heat sink. For instance, pure copper weighs 8.96 grams per cubic centimeter (g/cm3) and pure aluminum weighs 2.70 g/cm3. In many applications, several heat sinks need to be arrayed on a circuit board to dissipate heat from a variety of components on the board. If metallic heat sinks are employed, the sheer weight of the metal on the board can increase the chances of the board cracking or of other undesirable effects, and increases the weight of the component itself. Many metals do not exhibit a high surface thermal emissivity and thus do not effectively dissipate heat through the radiation mechanism.
Thus, there is a strong need for a non-metallic heat sink system effective for dissipating heat produced by a heat source such as a CPU. The heat sink system should exhibit a higher thermal conductivity and/or a higher thermal conductivity-to-weight ratio as compared to metallic heat sinks. These heat sinks must also be mass-producible, preferably using a cost-effective process. This processing ease requirement is important since metallic heat sinks can be readily produced in large quantities using scalable techniques such as extrusion, stamping, and die casting.
One group of materials potentially suitable for both EMI shielding and heat sink applications is the graphitic carbon or graphite. Carbon is known to have five unique crystalline structures, including diamond, fullerene (0-D nano graphitic material), carbon nano-tube or carbon nano-fiber (1-D nano graphitic material), graphene (2-D nano graphitic material), and graphite (3-D graphitic material). The carbon nano-tube (CNT) refers to a tubular structure grown with a single wall or multi-wall. Carbon nano-tubes (CNTs) and carbon nano-fibers (CNFs) have a diameter on the order of a few nanometers to a few hundred nanometers. Their longitudinal, hollow structures impart unique mechanical, electrical and chemical properties to the material. The CNT or CNF is a one-dimensional nano carbon or 1-D nano graphite material.
Bulk natural graphite powder is a 3-D graphitic material with each graphite particle being composed of multiple grains (a grain being a graphite single crystal or crystallite) with grain boundaries (amorphous or defect zones) demarcating neighboring graphite single crystals. Each grain is composed of multiple graphene planes that are oriented parallel to one another. A graphene plane in a graphite crystallite is composed of carbon atoms occupying a two-dimensional, hexagonal lattice. In a given grain or single crystal, the graphene planes are stacked and bonded via van der Waal forces in the crystallographic c-direction (perpendicular to the graphene plane or basal plane). Although all the graphene planes in one grain are parallel to one another, typically the graphene planes in one grain and the graphene planes in an adjacent grain are different in orientation. In other words, the orientations of the various grains in a graphite particle typically differ from one grain to another.
A graphite single crystal (crystallite) per se is anisotropic with a property measured along a direction in the basal plane (crystallographic a- or b-axis direction) being dramatically different than if measured along the crystallographic c-axis direction (thickness direction). For instance, the thermal conductivity of a graphite single crystal can be up to approximately 1,920 W/mK (theoretical) or 1,800 W/mK (experimental) in the basal plane (crystallographic a- and b-axis directions), but that along the crystallographic c-axis direction is less than 10 W/mK (typically less than 5 W/mK). Consequently, a natural graphite particle composed of multiple grains of different orientations exhibits an average property between these two extremes.
The constituent graphene planes of a graphite crystallite can be exfoliated and extracted or isolated from a graphite crystallite to obtain individual graphene sheets of carbon atoms provided the inter-planar van der Waals forces can be overcome. An isolated, individual graphene sheet of carbon atoms is commonly referred to as single-layer graphene. A stack of multiple graphene planes bonded through van der Waals forces in the thickness direction with an inter-graphene plane spacing of 0.3354 nm is commonly referred to as a multi-layer graphene. A multi-layer graphene platelet has up to 300 layers of graphene planes (<100 nm in thickness), but more typically up to 30 graphene planes (<10 nm in thickness), even more typically up to 20 graphene planes (<7 nm in thickness), and most typically up to 10 graphene planes (commonly referred to as few-layer graphene in scientific community). Single-layer graphene and multi-layer graphene sheets are collectively called “nano graphene platelets” (NGPs). Graphene sheets/platelets or NGPs are a new class of carbon nano material (a 2-D nano carbon) that is distinct from the 0-D fullerene, the 1-D CNT, and the 3-D graphite.
NGPs are typically obtained by intercalating natural graphite particles with a strong acid and/or oxidizing agent to obtain a graphite intercalation compound (GIC) or graphite oxide (GO), as illustrated in FIG. 1(A) and FIG. 1(B). The presence of chemical species or functional groups in the interstitial spaces between graphene planes serves to increase the inter-graphene spacing (d002, as determined by X-ray diffraction), thereby significantly reducing the van der Waals forces that otherwise hold graphene planes together along the c-axis direction. The GIC or GO is most often produced by immersing natural graphite powder (20 in FIG. 1(A)) in a mixture of sulfuric acid, nitric acid (an oxidizing agent), and another oxidizing agent (e.g. potassium permanganate or sodium perchlorate). The resulting GIC (22) is actually some type of graphite oxide (GO) particles. This GIC is then repeatedly washed and rinsed in water to remove excess acids, resulting in a graphite oxide suspension or dispersion, which contains discrete and visually discernible graphite oxide particles dispersed in water. There are two processing routes to follow after this rinsing step:
Route 1 involves removing water from the suspension to obtain “expandable graphite,” which is essentially a mass of dried GIC or dried graphite oxide particles. Upon exposure of expandable graphite to a temperature in the range of typically 800-1,050° C. for approximately 30 seconds to 2 minutes, the GIC undergoes a rapid expansion by a factor of 30-300 to form “graphite worms” (24), which are each a collection of exfoliated, but largely un-separated graphite flakes that remain interconnected.
In Route 1A, these graphite worms (exfoliated graphite or “networks of interconnected/non-separated graphite flakes”) can be re-compressed to obtain flexible graphite sheets or foils (26) that typically have a thickness in the range of 0.1 mm (100 μm)-0.5 mm (500 μm). Alternatively, one may choose to use a low-intensity air jet mill or shearing machine to simply break up the graphite worms for the purpose of producing the so-called “expanded graphite flakes” which contain mostly graphite flakes or platelets thicker than 100 nm (hence, not a nano material by definition).
Exfoliated graphite worms, expanded graphite flakes, and the recompressed mass of graphite worms (commonly referred to as flexible graphite sheet or flexible graphite foil) are all 3-D graphitic materials that are fundamentally different and patently distinct from either the 1-D nano carbon material (CNT or CNF) or the 2-D nano carbon material (graphene sheets or platelets, NGPs). Flexible graphite (FG) foils can be used as a heat spreader material, but exhibiting a maximum in-plane thermal conductivity of typically less than 500 W/mK (more typically <300 W/mK) and in-plane electrical conductivity no greater than 1,500 S/cm. These low conductivity values are a direct result of the many defects, wrinkled or folded graphite flakes, interruptions or gaps between graphite flakes, and non-parallel flakes (e.g. SEM image in FIG. 2(B)). Many flakes are inclined with respect to one another at a very large angle (e.g. mis-orientation of 20-40 degrees).
In Route 1B, the exfoliated graphite is subjected to high-intensity mechanical shearing (e.g. using an ultrasonicator, high-shear mixer, high-intensity air jet mill, or high-energy ball mill) to form separated single-layer and multi-layer graphene sheets (collectively called NGPs, 33). Single-layer graphene can be as thin as 0.34 nm, while multi-layer graphene can have a thickness up to 100 nm. In the present application, the thickness of multi-layer NGPs is typically less than 20 nm.
Route 2 entails ultrasonicating the graphite oxide suspension for the purpose of separating/isolating individual graphene oxide sheets from graphite oxide particles. This is based on the notion that the inter-graphene plane separation bas been increased from 0.3354 nm in natural graphite to 0.6-1.1 nm in highly oxidized graphite oxide, significantly weakening the van der Waals forces that hold neighboring planes together. Ultrasonic power can be sufficient to further separate graphene plane sheets to form separated, isolated, or discrete graphene oxide (GO) sheets. These graphene oxide sheets can then be chemically or thermally reduced to obtain “reduced graphene oxides” (RGO) typically having an oxygen content of 0.001%-10% by weight, more typically 0.01%-5% by weight.
For the purpose of defining the claims of the instant application, NGPs include discrete sheets/platelets of single-layer and multi-layer versions of graphene, graphene oxide, or reduced graphene oxide with an oxygen content of 0-10% by weight, more typically 0-5% by weight, and preferably 0-2% by weight. Pristine graphene has essentially 0% oxygen. Graphene oxide (including RGO) typically has 0.001%-46% by weight of oxygen.
Flexible graphite foils may be obtained by compressing or roll-pressing exfoliated graphite worms into paper-like sheets. For electronic device thermal management applications (e.g. as a heat sink material), flexible graphite (FG) foils have the following major deficiencies:                (1) As indicated earlier, FG foils exhibit a relatively low thermal conductivity, typically <500 W/mK and more typically <300 W/mK. By impregnating the exfoliated graphite with a resin, the resulting composite exhibits an even lower thermal conductivity (typically <<200 W/mK, more typically <100 W/mK).        (2) Flexible graphite foils, without a resin impregnated therein or coated thereon, are of low strength, low rigidity, and poor structural integrity. The high tendency for flexible graphite foils to get torn apart makes them difficult to handle in the process of making a heat sink. As a matter of fact, the flexible graphite sheets (typically 50-200 μm thick) are so “flexible” that they are not sufficiently rigid to make a fin component material for a finned heat sink.        (3) Another very subtle, largely ignored or overlooked, but critically important feature of FG foils is their high tendency to get flaky with graphite flakes easily coming off from FG sheet surfaces and emitting out to other parts of a microelectronic device. These highly electrically conducting flakes (typically 1-200 μm in lateral dimensions and >100 nm in thickness) can cause internal shorting and failure of electronic devices.        
Similarly, solid NGPs (including discrete sheets/platelets of pristine graphene, GO, and GRO), when packed into a film, membrane, or paper sheet (34) of non-woven aggregates, typically do not exhibit a high thermal conductivity unless these sheets/platelets are closely packed and the film/membrane/paper is ultra-thin (e.g. <1 μm, which is mechanically weak). This is reported in our earlier U.S. patent application Ser. No. 11/784,606 (Apr. 9, 2007). In general, a paper-like structure or mat made from platelets of graphene, GO, or RGO (e.g. those paper sheets prepared by vacuum-assisted filtration process) exhibit many defects, wrinkled or folded graphene sheets, interruptions or gaps between platelets, and non-parallel platelets (e.g. SEM image in FIG. 3(B)), leading to relatively poor thermal conductivity, low electric conductivity, and low structural strength. These papers or aggregates of discrete NGP, GO or RGO platelets alone (without a resin binder) also have a tendency to get flaky, emitting conductive particles into air.
Our earlier application (U.S. application Ser. No. 11/784,606) also disclosed a mat, film, or paper of NGPs infiltrated with a metal, glass, ceramic, resin, and CVD carbon matrix material (graphene sheets/platelets being the filler or reinforcement phase, not the matrix phase in this earlier application). Haddon, et al. (US Pub. No. 2010/0140792, Jun. 10, 2010) also reported NGP thin film and NGP-reinforced polymer matrix composites for thermal management applications. The NGP-reinforced polymer matrix composites, as an intended thermal interface material, have very low thermal conductivity, typically <<2 W/mK. The NGP films of Haddon, et al are essentially non-woven aggregates of discrete graphene platelets, identical to those of our earlier invention (U.S. application Ser. No. 11/784,606). Again, these aggregates have a great tendency to have graphite particles flaking and separated from the film surface, creating internal shorting problem for the electronic device containing these aggregates. They also exhibit low thermal conductivity unless made into thin films (10 nm-300 nm, as reported by Haddon, et al) which are very difficult to handle in a real device manufacturing environment. Balandin, et al (US Pub. No. 2010/0085713, Apr. 8, 2010) disclosed a graphene layer produced by CVD deposition or diamond conversion for heat spreader application. More recently, Kim, et al (N. P. Kim and J. P. Huang, “Graphene Nanoplatelet Metal Matrix,” US Pub. No. 2011/0108978, May 10, 2011) reported metal matrix infiltrated NGPs. However, the metal matrix is too heavy and the resulting metal matrix composite does not exhibit a high thermal conductivity.
Another prior art material for thermal management or EMI shielding application is the pyrolitic graphite film. The lower portion of FIG. 1(A) illustrates a typical process for producing prior art pyrolitic graphite films from a polymer. The process begins with carbonizing a polymer film 46 at a carbonization temperature of 400-1,500° C. under a typical pressure of 10-15 Kg/cm2 for 2-10 hours to obtain a carbonized material 48, which is followed by a graphitization treatment at 2,500-3,200° C. under an ultrahigh pressure of 100-300 Kg/cm2 for 1-5 hours to form a graphitic film 50. There are several major drawbacks associated with this process for producing graphitic films:    (1) Technically, it is utmost challenging to maintain such an ultrahigh pressure (>100 Kg/cm2) at such an ultrahigh temperature (>2,500° C.). The combined high temperature and high pressure conditions, even if achievable, are not cost-effective.    (2) This is a difficult, slow, tedious, energy-intensive, and very expensive process.    (3) This polymer graphitization process is not conducive to the production of either thick graphitic films (>50 μm) or very thin films (<10 μm).    (4) In general, high-quality graphitic films could not be produced with a temperature lower than 2,700° C., unless when a highly oriented polymer is used as a starting material (e.g. please see Y. Nishikawa, et al. “Filmy graphite and process for producing the same,” U.S. Pat. No. 7,758,842 (Jul. 20, 2010)) or a catalytic metal is brought in contact with a highly oriented polymer during carbonization and graphitization (Y. Nishikawa, et al. “Process for producing graphite film,” U.S. Pat. No. 8,105,565 (Jan. 31, 2012)). This high degree of molecular orientation, as expressed in terms of optical birefringence, is not always possible to achieve. Further, the use of a catalytic metal tends to contaminate the resulting graphite films with metallic elements.    (5) The resulting graphitic films tend to be brittle and of low mechanical strength.
A second type of pyrolytic graphite is produced by high temperature decomposition of hydrocarbon gases in vacuum followed by deposition of the carbon atoms to a substrate surface. This vapor phase condensation of cracked hydrocarbons is essentially a chemical vapor deposition (CVD) process. In particular, highly oriented pyrolytic graphite (HOPG) is the material produced by the application of uniaxial pressure on deposited pyrocarbon or pyrolytic graphite at very high temperatures (typically 3,000-3,300° C.). This entails a thermo-mechanical treatment of combined mechanical compression and ultra-high temperature for an extended period of time in a protective atmosphere; a very expensive, energy-intensive, and technically challenging process. The process requires ultra-high temperature equipment (with high vacuum, high pressure, or high compression provision) that is not only very expensive to make but also very expensive and difficult to maintain.
It is an object of the present invention to provide a process for producing graphitic films that exhibit a combination of exceptional thermal conductivity, electrical conductivity, and mechanical strength unmatched by any material of comparable thickness range.
Another object of the present invention is to provide a cost-effective process for producing a thermally conductive graphitic film from a nano graphene platelet-filled polymer or other type of carbon precursor material (e.g. pitch, monomer, oligomer, organic substance, such as maleic acid) through controlled carbonization and graphitization.
In particular, the present invention provides a process capable of producing a graphitic film from a nano graphene platelet-filled polymer or other carbon precursor material at a carbonization temperature and/or a graphitization temperature lower than the carbonization temperature and/or a graphitization temperature required of successfully producing a graphitic film of a comparable conductivity from a corresponding neat polymer alone (without nano graphene platelet, NGP).
As compared to conventional processes, this inventive process involves significantly lower heat treatment temperatures, shorter heat treatment times and lower amount of energy consumed, resulting in graphitic films that are of higher thermal conductivity, higher electrical conductivity, and higher strength.